Vitamin B12, also known as cobalamin, is a water-soluble B-vitamin [1
]. Vitamin B12 is synthesized by bacteria, and finds its way into the human food supply predominantly via animal products [2
]. This places low consumers of animal products—vegetarians, and particularly vegans—at risk of low intake [3
] and metabolic deficiency [4
]. Other causes of deficiency are caused by malabsorption of the vitamin due to tropical sprue [5
], Helicobacter pylori
], hypochlorhydria [7
], and gastric atrophy [8
]. Human deficiencies of vitamin B12 result in adverse effects to the individual that, when prevalent in the population, may be of public health significance [9
]. The clinical manifestations of vitamin B12 deficiency result in haematologic, neurologic, psychiatric, and other chronic illnesses [10
]. Breastfeeding infants of vitamin B12-deficient mothers are at particular risk of potentially life-threatening complications if the deficiency is not identified and rectified [11
]. The prevalence of frank deficiency resulting in haematological or neurological problems is relatively uncommon, but subclinical deficiency has been estimated to range from 2.5% to 26% in various populations [12
]. Functional markers of vitamin B12 status, namely methylmalonic acid and homocysteine, are elevated when vitamin B12 status is deficient (<148 pmol/L) or suboptimal (<221 pmol/L) [12
]. The functional biomarkers may also be raised when vitamin B12 status is between 148 and 221 pmol/L. This sub-optimal status may confer additional risk for neurodegenerative disease [13
] and for women, neural tube defects in their offspring [15
Areas identified as having populations at risk of low vitamin B12 intakes include the Indian subcontinent [16
], Central and Southern America, and parts of Africa [17
]. Whether people who emigrate from these areas are similarly susceptible to low vitamin B12 intakes will be of interest to their new country. People of Indian and Pakistani descent living in Toronto were found to have a high prevalence of vitamin B12 deficiency, attributed to a low intake of animal products, compared with the deficiency rates in the general population [18
]. New Zealand has immigrant populations from across Asia, with both vitamin B12 intakes and biochemical data available for analysis from the New Zealand Adult Nutrition Survey, a survey conducted on a nationally representative sample in 2008/2009. Given the previous research, we hypothesize that people of South Asian origin living in New Zealand will have lower vitamin B12 intakes and status compared with other immigrant and indigenous groups. We further hypothesize that cultural and religious practices will predispose South Asian people to low intakes of animal products. Data collected in 1997 from a previous national nutrition survey suggest that a high proportion of older New Zealanders were at risk of vitamin B12 insufficiency among a predominantly European sample of adults [19
]. Analysis of data from the current New Zealand Adult Nutrition Survey provides an opportunity to assess vitamin B12 intakes and determinants of biochemical status among a unique set of ethnic groups across a wider age range. Thus, the primary aims of the study are to compare vitamin B12 intake and status among ethnic groups living in New Zealand, with a secondary objective to assess dietary factors associated with intake and status.
2. Materials and Methods
The New Zealand Adult Nutrition Survey was a nationally representative, cross-sectional survey of 4721 New Zealanders aged 15 years and above. The New Zealand Health and Disability Multi-Region Ethics Committee granted ethical approval for the survey (MEC/08/04/049). A full account of the methodology has been published [20
]. In brief, participants were recruited using a three-stage, stratified, area-based sampling frame involving 607 geographical areas, followed by the selection of households within each area and a randomly selected respondent within each household. Informed, written consent was obtained from the participants. The survey was undertaken from October 2008 to October 2009 and had a response rate of 61%.
2.1. Data Collection: Dietary, Anthropometric, and Lifestyle
During a home visit, a trained interviewer assessed weight, height, and waist circumference using calibrated instruments, and collected socio-demographic and dietary information using proprietary computer-assisted interview software. An indicator of socioeconomic status of participants was attained via the New Zealand Index of Deprivation 2006 [21
]. Dietary information was via a four-stage multi-pass 24-h diet recall and a dietary habits questionnaire. The four stages of the 24-h recall were as follows:
A “quick list” of all foods, beverages, and dietary supplements consumed during the preceding day (midnight to midnight) was obtained.
Detailed descriptions were obtained of all items consumed, using specific questions and prompts, including cooking method, recipe for mixed dishes (where known), any additions made before consumption, brand and product name, time consumed, and where the food was sourced. Brand and product name were determined using a bar code scanner for food items where the composition was brand specific and packaging was available.
Estimates were made of amounts of items consumed, wherever possible (e.g., cups, tablespoons), using food photographs, shape dimensions, food portion assessment aids (e.g., dried beans), and packaging information.
All items were reviewed in chronological order. Any additions and changes were made at this point.
Information on the nutrient content of foods was obtained from the New Zealand Food Composition Database [22
] and nutrient composition calculated using bespoke software developed in-house by the University of Otago Department of Human Nutrition staff [23
The following four questions were used from the Adult Nutrition Survey Dietary Habits Questionnaire [24
In the past four weeks which of the following have you eaten at all?
Red meat—such as beef, pork, mutton, lamb and goat.
Chicken—such as chicken breast, drumsticks, or whole chickens, but not chicken nuggets or chicken roll; Processed meats—such as ham, bacon, sausages, luncheon, canned corned beef, pastrami and salami; Seafood—such as fish or shellfish
Responses to these questions were coded Selected (1), Not selected (0).
On average how many servings of fruit—fresh, frozen, canned, or stewed—do you eat per day?
Never, I don’t eat fruit; Less than one serving per day; 1 serving; 2 servings; 3 servings; 4 or more servings; Do not know.
On average how many servings of vegetables—fresh, frozen, or canned—do you eat per day?
Never, I don’t eat vegetables; Less than one serving per day; 1 serving; 2 servings; 3 servings; 4 or more servings; Do not know.
What type of milk do you use the most of?
None, I don’t use milk; Whole or standard milk; Reduced fat; Skim or Trim; Soy milk; Other (such as rice, goats milk); Do not know.
The responses to questions 2 and 3 (fruit and vegetables) were summarized and analysed as a single “fruit and vegetable” variable to assess whether participants were meeting the “5+ a day” recommendation.
The responses to question 4 were dichotomized into cow’s milk drinkers or non-cow’s milk drinkers, to assess the specific contribution of vitamin B12 to the diet from cow’s milk.
A question on smoking (SC2_Q11: How often do you currently smoke?) and one on alcohol intake (SC2_Q16: How often do you have a drink containing alcohol?) were also used in the analysis.
2.2. Blood Biochemistry
A non-fasting blood sample was obtained from 3348 non-pregnant survey respondents, giving a response rate of 71% of the total sample. To give a blood sample, participants attended a health clinic in their locality. The blood samples were processed by the local clinics, and aliquots of serum were sent overnight at 4 °C to the University of Otago, where the samples were stored at −80 °C until analysis. There was insufficient blood to complete a vitamin B12 analysis for 276 participants. Vitamin B12 analyses were performed using a chemiluminescent kit (Roche Diagnostics GmbH, Mannheim, Germany) on a Roche Hitachi Elecsys2010 system (Hitachi High-Technologies Corporation, Tokyo, Japan). The within-batch coefficients of variation of a pooled plasma sample and the manufacturer’s control were both <5%. The absolute concentrations of repeated control samples were within 3% of the manufacturer’s stated concentration.
Ethnicity was self-reported, using the Statistics New Zealand standard ethnicity question from the New Zealand Census 2006. The question was “Which ethnic group do you belong to?”, with tick boxes for New Zealand European, Māori, Samoan, Cook Island Māori, Tongan, Niuean, Chinese, Indian, and Other. People could identify with more than one ethnicity, so for the purpose of this analysis, a single ethnicity was assigned based on pre-determined prioritization. For example, a person was classified as Māori if any one of their recorded ethnicities was Māori. The prioritization order was Māori followed by Pacific Island, which included those people who identified as Samoan, Cook Island Māori, Tongan, Niuean, and Fijian. Other ethnicities were encompassed within the term New Zealand European and Other. South Asian refers to all people who have ancestral origins in the Indian subcontinent (the countries of India, Afghanistan, Pakistan, Sri Lanka, Nepal, Bangladesh, Bhutan, and the Maldives). Fijian Indians were also included in the South Asian category, because their ancestry is from the Indian subcontinent, and it is hypothesized that they brought their cultural habits with them to the new country of residence (the South Asian diaspora) [25
]. East and South-East Asian included Chinese, Malaysian Chinese, Taiwanese, Filipino, Cambodian, Vietnamese, Burmese, Indonesian, Japanese, Korean, and Tibetan people. There were 49 people who did not fit into any of the ethnic groupings; these were predominantly of African origin, and their data were not used in the analyses presented here.
Survey-prioritized weighting was used in all analyses to ensure that no group was under- or over-represented, such that the estimates of means, percentiles, and proportions were representative of the total resident adult population of New Zealand. Blood-prioritized weight was exclusively used when performing statistical analyses of serum vitamin B12 data. The procedure for weighting is described in detail in the Survey Methodology Report [20
Multiple linear regression analysis was used to quantify the independent effect of dietary, lifestyle, and socioeconomic exposures on serum vitamin B12. Age, sex, and ethnicity were regarded as confounding variables and were controlled for in the analyses. Pre-specified sub-groupings included grouping by ethnicity and sex. The adult estimated average B12 requirement of 2 µg/day was used to determine adequacy of vitamin B12 intake [26
]. The cut-off used to define vitamin B12 deficiency was <148 pmol/L (200 pg/mL), and depletion was defined as 148–221 pmol/L (200–300 pg/mL), as suggested in the literature [27
]. There were some exceptionally high vitamin B12 intakes and serum values in the dataset. Extreme values were quantified using an outlier algorithm, in which the inter-quartile range was multiplied by 1.5 and added to the third quartile [28
]. Using this algorithm, the serum and intake outlier values calculated were 701 pmol/L and 9.56 µg/day, respectively. Intakes of around 10 µg/day (five times the estimated average requirement) and serum concentrations in excess of 700 pmol/L were more likely the consequence of supplementation or injection than diet alone. As high intakes from supplements and or injections were secondary to the aim, only serum vitamin B12 concentrations values <701 pmol/L and vitamin B12 intakes <9.56 µg/day were included in the regression, resulting in the data of 12 people being dropped. In total, the serum vitamin B12 concentrations of 3011 participants were included in this analysis.
Variables that were not normally distributed were log transformed. Chi-squared and Fisher’s exact tests were used to check for association between categorical variables. Differences among ethnic groups were assessed using one-way ANOVA with Bonferroni multiple-comparison test. Statistical analysis was performed with STATA version 12 (StataCorp LP, College Station, TX, USA).
Our data indicate variability in the vitamin B12 intakes and biochemical status among residents of New Zealand based on regional ethnic groupings. East and South-East Asians had a similar intake and status to Māori and Pacific, with Māori and Pacific having a higher vitamin B12 intake and status compared with New Zealand Europeans. South Asians had intakes lower than East and South-East Asians and Māori and Pacific, with serum concentrations lower than all other groups. Thus, Māori and Pacific were the least likely to have inadequate intakes, compared with New Zealand Europeans, and the latter group was more likely to have an adequate biochemical status than the South Asians.
These data are consistent with other work. In Canada, a high prevalence of vitamin B12 deficiency was found among South Asian men and women registered with a Toronto clinic [18
]. In the United Kingdom, healthy European men had higher vitamin B12 status (357 pmol/L) compared with South Asian counterparts (270 pmol/L) [29
]. That value of 270 pmol/L is very comparable to the mean serum concentration of 282 pmol/L in our South Asian group living in New Zealand. A smaller proportion of our South Asian group were above the sufficient cutoff than the other groups, indicating that as a group, South Asians may be more at risk of subclinical deficiency, based on elevated functional markers of vitamin B12 metabolism [30
]. Limitations of our work were the absence of such functional biomarkers of vitamin B12 status, namely total plasma homocysteine, methylmalonic acid, and holo-transcobalamin. These functional biomarkers are useful in assessing the likelihood of depletion in the intermediate range of serum vitamin B12 concentration (148–221 pmol/L) [27
]. Plasma total homocysteine concentration is also affected by folate status via folate–cobalamin interactions [31
]. However, serum vitamin B12 concentration in and of itself is a valid indicator of vitamin B12 status, with a value of 148 pmol/L widely used in the epidemiologic setting to indicate insufficiency [32
]. Around 3% of the South Asian and New Zealand European groups were below this cutoff. Overt vitamin B12 deficiency has been associated with megaloblastic anemia in Asian vegetarians [33
]. The elderly may also be compromised, with vitamin B12 being inversely associated with cognitive tests of spatial copying skills [34
]. Pregnant women with vitamin B12 deficiency and insufficiency are at risk of having a child with birth defects, and low maternal status further puts the infant at risk of frank deficiency and failure to thrive, particularly if the mother breastfeeds [35
]. If the low maternal vitamin B12 status is due to an inadequate intake, perhaps for cultural or religious reasons, then it is likely that the child will be raised in a low vitamin B12 environment with longer-term consequences. Low vitamin B12 status in Indian mothers has been associated with poorer scores of cognitive tests in their nine-year-old children [36
]. Thus, lower mean vitamin B12 status exposes a higher proportion of South Asian individuals, compared with other ethnic groups, to an elevated risk of metabolic insufficiency, with around 3% of South Asians and New Zealand Europeans at risk of clinical deficiency.
The relatively low vitamin B12 status of South Asians has been attributed to vegetarianism [29
]. However, low intakes and status are not exclusively associated with strict vegetarianism, as South Asians may avoid certain meats or consume limited amounts of animal products, due to cultural and religious practices [37
]. We did not ask directly about cultural and religious practices, but the relatively high proportion who avoided meat suggested that such practices may be contributing factors to the lower vitamin B12 status of the South Asian group. Indeed, meat eaters had considerably higher serum vitamin B12 concentrations than non-meat eaters, resulting in meat intake being a major predictor of vitamin B12 serum concentrations. A major source of dietary vitamin B12 for South Asians was milk, with 98% of the group consuming it. Despite this uptake, overall intake of vitamin B12 tended to be lower, and biochemical status was lower in the South Asians than the other ethnic groups. Vitamin B12 bioavailability from milk is good unless heated, a process that destroys considerable amounts of the vitamin [2
]. This may be a problem for South Asians, as there are reports of widespread heating of milk in Indian cuisine [38
]. The last dietary factor to be assessed was fruit and vegetable intake. Given that fruits and vegetables do not contain vitamin B12, we tested whether meeting the guidelines of five or more servings a day negatively impacted serum vitamin B12 concentrations, hypothesizing that a large intake of fruit and vegetables could displace vitamin B12-containing animal products. This did not appear to be the case, as in multiple regression, fruits and vegetables were not predictive of serum vitamin B12 concentrations. This is consistent with a lack of correlation between fruit and vegetable intake and serum vitamin B12 status in adult women [39
]. Our data complement those presented previously obtained for the entire New Zealand Adult Nutrition Survey sample, for which milk was reported to be the largest single contributor of vitamin B12 intakes when major food groups were sub-classified into their components [40
]. For example, dairy foods were sub-classified into milk, cheese, and other dairy products, while meat was sub-classified into beef, poultry, lamb, pork, etc. Despite the prominence of milk as a contributor of vitamin B12 intake in the overall sample, the strongest predictors of serum vitamin B12 concentration in our analyses were ethnic groupings and meat consumption (combined intake of all types).
A particular strength of this study was the use of robust dietary assessment methodology coupled with a dietary habits questionnaire, which provided insight into the dietary factors associated with serum vitamin B12 status. How generalizable these findings are to the wider South Asian community living in New Zealand is uncertain, because we did not ascertain the length of time participants had been in the country, or indeed whether they were first- or subsequent-generation residents. The length of residence could influence vitamin B12 status, because a migratory effect has been found in which Indians living in the United Kingdom had higher vitamin B12 status than a matched cohort living in their home villages in India [41
]. Nevertheless, the vitamin B12 status of the United Kingdom dwellers was still relatively low, at 240 pmol/L. Another limitation was the small sample size of the South Asian group, particularly of those in the older age brackets. As vitamin B12 status tended to decline with advancing age, an underrepresentation of older South Asians may have produced an underestimate of the overall prevalence of vitamin B12 deficiency in this population subgroup. Nevertheless, our data are consistent with a wider body of evidence indicating that South Asians living in various countries are at risk of vitamin B12 insufficiency or deficiency. It would be helpful for health agencies around the world to be aware of this potential for compromised vitamin B12 status in migrant South Asian communities, particularly in the identification of individuals in at-risk groups, such as the elderly and young women capable of childbearing, as well as those who do not eat meat.